Kalman filtering based on
the maximum correntropy criterion
in the presence of non-Gaussian noise

 

Reza Izanloo, Seyed Abolfazl Fakoorian, Hadi Sadoghi, and Dan Simon

This paper deals with state estimation in the presence of non-Gaussian noise. Since the Kalman filter uses only second-order signal information, it is not optimal in non-Gaussian noise environments. The maximum correntropy criterion (MCC) is a new approach to measure the similarity of two random variables. MCC uses information from the higher-order statistics of the signals. The correntropy filter (C-Filter) uses the properties of the MCC for state estimation. This paper first improves the performance of the C-filter by modifying its derivation to obtain a modified correntropy filter (MC-Filter). This paper then uses the properties of MCC and weighted least squares (WLS) to propose an MCC filter in Kalman filter form which we call the MCC-KF. Simulation results show the superiority of the MCC-KF compared with the C-Filter, the MC-Filter, the unscented Kalman filter, the ensemble Kalman filter, and the Gaussian sum filter, in the presence of two different types of non-Gaussian disturbances (shot noise and Gaussian mixture noise).

The MATLAB software that was used to derive the results in the paper can be downloaded in this zip file (start with the "README.TXT" file).

 

Reference

 

R. Izanloo, S. Fakoorian, H. Sadoghi, and D. Simon, "Kalman filtering based on the maximum correntropy criterion in the presence of non-Gaussian noise," 50th Annual Conference on Information Systems and Sciences, March 2016 - pdf, 292 KB

 


Professor Simon's Home Page

 

Department of Electrical Engineering and Computer Science

 

Cleveland State University

 


Last Revised: February 24, 2016